Table of Contents

Getting Started

Installation

Install FlameCsv using the NuGet package manager:

dotnet add package FlameCsv

Basic Usage

Reading CSV data is as simple as:

// Reading from a string
IEnumerable<User> users = Csv.From("id,name\n1,John\n2,Jane").Read<User>();

// Reading from a file
await foreach (User user in Csv.FromFile("users.csv").ReadAsync<User>(cancellationToken))
{
    Console.WriteLine(user.Name);
}

// Processing records in parallel
await Csv.From(x)
    .AsParallel(cancellationToken)
    .ForEachAsync<Obj>(
        async (ArraySegment<Obj> records, CancellationToken ct) =>
        {
            await ProcessBatch(records, ct);
        }
    );

Writing CSV is just as easy:

var users = new[]
{
    new User { Id = 1, Name = "John" },
    new User { Id = 2, Name = "Jane" }
};

// Writing to a string
StringBuilder destination = new();
Csv.To(destination).Write(users);
return destination.ToString();

// Writing to a file
await Csv.ToFile("users.csv").WriteAsync(users, cancellationToken);

// Write streamed in parallel
await Csv.To(stream)
    .WithUtf8Encoding()
    .AsParallel(new CsvParallelOptions{ChunkSize = 256, CancellationToken = cancellationToken})
    .WriteUnorderedAsync(
        GetUsersAsIAsyncEnumerable(),
        CsvOptions<char>.Default
    );

Configuration

FlameCsv is highly configurable. Common options include:

Example of custom configuration:

CsvOptions<char> options = new()
{
    Delimiter = ';',
    Newline = CsvNewline.Platform, // CRLF or LF depending on OS
    Trimming = CsvFieldTrimming.Leading,
    HasHeader = true,
    NormalizeHeader = static h => h.Trim().ToString().Replace("_", ""),
};

The configuration object is identical for char and byte, and UTF16 <-> UTF8 conversion is handled automatically. Only differences are the converters, which are separate for char and byte.

Next Steps

Edit this page